Founder StoryApril 2026 · 7 min read
Mapping GPUs to LLMs (and back): A bandwidth-based estimator for local inference
How a weekend project to stop juggling GPU specs, model cards, and Reddit threads turned into a physics-based recommendation engine.
Read post